Is sweet potato good for weight loss?

Not as far as any trial shows. We found no trial that gives people whole sweet potato and compares it with another food of the same calories. The closest is an 8-week randomized trial in 58 overweight office workers in Taiwan, where a white sweet potato meal replacement of 132 g a day went with about a 5 percent drop in body weight. The comparison group ate a normal diet rather than a matched meal replacement, and it lost weight too, so the trial cannot pin the loss on the sweet potato.

Unsettled. Sweet potato is a starchy food with no tested weight effect of its own. The two kinds of human trial that exist used a commercial meal replacement formula or a white sweet potato extract taken by people with diabetes. In both, the control group lost weight as well.

What the trials found

The meal replacement trial

Fifty-eight overweight white-collar workers were randomized for eight weeks. The group on the white sweet potato meal replacement showed a 5 percent decrease in body weight, body fat, body mass index and mid-arm circumference, and a 3.5 percent decrease in glycated hemoglobin. The normal-diet group also fell significantly on these measures. Because the comparison group got no meal replacement, this design cannot separate the effect of replacing meals from the effect of sweet potato.

The diabetes extract trials

Caiapo is an extract of white sweet potato, a different product from the vegetable. In a 12-week randomized trial of 61 people with type 2 diabetes taking 4 g a day, body weight fell in both the extract and the placebo groups. The authors put that down to a better controlled lifestyle. The abstract does not report the difference between the groups.

A Cochrane review found only three small randomized trials of this extract, in 140 people, all from one trialist, with risk of bias unclear or high. Weight was not a pooled outcome. What it did pool was blood sugar: glycated hemoglobin fell by 0.3 percent at three to five months on 4 g a day, in two trials with 122 participants, with a confidence interval from 0.6 to 0.04. A separate meta-analysis of single herbs in type 2 diabetes reached the same 0.30 percent from the same small set of trials. That is a modest effect of an extract on blood sugar in people with diabetes, and it says nothing about weight.

Cooking changes the blood sugar response

The glycemic index of sweet potato depends heavily on how it is cooked. In ten healthy adults eating ten Jamaican cultivars, boiled sweet potato came out between 41 and 50, while baked ran from 82 to 94 and roasted from 79 to 93.

That is a real and large difference in how fast blood sugar rises. It is not evidence about weight, and no trial has linked the glycemic index of sweet potato to body weight.

Weight loss claims for purple sweet potato come from studies in rats. We found no human trial.

Who should be careful

Not for everyone. Eating a lot of orange sweet potato for a long time can turn the skin yellow-orange. NIH calls this carotenodermia and describes it as harmless, and it fades when intake drops. The lung cancer warning in the same NIH passage concerns high-dose beta-carotene supplements in smokers. Food is outside it.

Beyond skin color, this page has no cautions to report. We found no evidence on sweet potato and weight that concerns pregnancy, medicines or allergy, and that is an absence of data rather than an all-clear.

What expert bodies say

We found no statement from a health authority on sweet potato and body weight. The one agency text on this page is the NIH note on carotenodermia above.

How we searched

What would change this answer

• A randomized trial of whole sweet potato against another starchy food of the same calories, with body weight as the outcome. That is the trial the question needs, and nothing like it was registered when we searched.
• A repeat of the meal replacement trial with a matched replacement in the control group. That would show whether the sweet potato or the replacement itself did the work.
• Human trials of purple sweet potato. The weight claims for it rest on rodents.
• A trial that tests whether eating sweet potato boiled rather than baked changes anything beyond the blood sugar curve.

Sources

1. ev-spwl-01 · Meta-analysis or systematic review · Cochrane systematic review and meta-analysis of RCTs, search to February 2013 · n = 140
   There was a statistically significant improvement in glycosylated haemoglobin A1c (HbA1c) at three to five months with 4 g/day sweet potato preparation compared to placebo (mean difference -0.3% (95% confidence interval -0.6 to -0.04); P = 0.02; 122 participants; 2 trials).
   
   Who: adults with type 2 diabetes in RCTs of sweet potato preparations vs placebo, 6 weeks to 5 months
   Effect: HbA1c MD -0.3% (95% CI -0.6 to -0.04, 2 trials, 122 participants) at 3 to 5 months; no serious adverse effects; risk of bias unclear or high
   Certainty: Very small evidence base from a single trialist; extract (Caiapo), not whole sweet potato; outside the PubMed snapshot.
   Ooi CP, Loke SC. Cochrane Database Syst Rev. 2013, CD009128 · checked 2026-10-07 · we read the abstract
2. ev-spwl-02 · Meta-analysis or systematic review · systematic review and meta-analysis of randomized placebo-controlled trials · n = 487 (all herbs)
   The pooled mean differences in HbA(1c) were -0.30% (95% CI -0.04% to -0.57%; P = 0.02), -1.92% (95% CI -0.51% to -3.32%; P = 0.008), and -1.13% (95% CI -0.11% to -2.14%; P = 0.03), respectively, for Ipomoea batatas, Silybum marianum, and Trigonella foenum-graecum.
   
   Who: patients with type 2 diabetes in placebo-controlled RCTs of single herbs, at least 8 weeks
   Effect: Ipomoea batatas: HbA1c MD -0.30% (95% CI -0.04 to -0.57), fasting glucose MD -10.20 mg/dL (-5.32 to -15.08)
   Certainty: Sweet potato subset drawn from the same small trial set as the Cochrane review; glycemic outcome, not weight.
   J Ethnopharmacol, 2011 · checked 2026-10-07 · we read the abstract
3. ev-spwl-03 · Randomized controlled trial(s) · multicentre randomized placebo-controlled trial, 12 weeks · n = 61
   A decrease in body weight was observed in both the placebo group (P = 0.0027) and in the Caiapo group (P < 0.0001), probably due to a better- controlled lifestyle.
   
   Who: type 2 diabetic patients treated by diet, mean BMI about 28
   Effect: weight decreased in placebo (P = 0.0027) and Caiapo (P < 0.0001) groups; HbA1c 7.21 to 6.68% on Caiapo vs 7.04 to 7.10% on placebo
   Certainty: Between-group weight difference not reported in the abstract.
   Diabetes Care, 2004 · checked 2026-10-07 · we read the abstract
4. ev-spwl-04 · Randomized controlled trial(s) · randomized controlled trial, 8 weeks, meal replacement vs normal diet · n = 58
   Moreover, the WSP-MR group demonstrated a 5% decrease in body weight, body fat, body mass index, and mid-arm circumference and a 3.5% decrease in glycated hemoglobin levels (p < 0.05).
   
   Who: overweight white-collar workers, Taiwan
   Effect: about 5% decrease in body weight, body fat, BMI and mid-arm circumference; HbA1c -3.5%; anthropometrics fell significantly in both groups
   Certainty: Control was a normal diet, not a calorie-matched meal replacement, so the effect cannot be put down to sweet potato itself; commercial formula.
   Nutrients, 2019 · checked 2026-10-07 · we read the abstract
5. ev-spwl-05 · Observational data · glycemic index feeding study, 10 cultivars and 4 cooking methods · n = 10
   Samples prepared by boiling had the lowest GI (41 ± 5-50 ± 3), while those processed by baking (82 ± 3-94 ± 3) and roasting (79 ± 4-93 ± 2) had the highest GI values.
   
   Who: nondiabetic adults (5 men, 5 women, mean age 27)
   Effect: GI boiled 41 to 50; baked 82 to 94; roasted 79 to 93; overall range 41 to 93
   Certainty: Small, single country cultivars; GI is a blood sugar measure, not a weight outcome; outside the PubMed snapshot.
   Bahado-Singh PS, et al. J Nutr Metab. 2011 · checked 2026-10-07 · we read the abstract
6. ev-spwl-06 · Position of an expert body · agency fact sheet · n = —
   The most common effect of long-term, excess beta-carotene is carotenodermia, a harmless condition in which the skin becomes yellow-orange [3].
   
   Who: general population with long-term high beta-carotene intake
   Effect: carotenodermia, harmless and reversible; beta-carotene not known to be teratogenic
   Certainty: Agency statement; the lung cancer risk in the same passage concerns high-dose supplements in smokers, not food.
   NIH Office of Dietary Supplements, Vitamin A and Carotenoids Fact Sheet for Health Professionals · checked 2026-10-07 · we read the section